Migrate GemliteUIntXWeightOnlyConfig to v2 design (UIntxWeightOnlyConfig)#4079
Closed
jerryzh168 wants to merge 1 commit intogh/jerryzh168/46/basefrom
Closed
Migrate GemliteUIntXWeightOnlyConfig to v2 design (UIntxWeightOnlyConfig)#4079jerryzh168 wants to merge 1 commit intogh/jerryzh168/46/basefrom
jerryzh168 wants to merge 1 commit intogh/jerryzh168/46/basefrom
Conversation
…fig) Replace the old AQT-based GemliteUIntXWeightOnlyConfig with a new v2 UIntxWeightOnlyConfig backed by UIntxBitPackedTensor(TorchAOBaseTensor), removing the dependency on AffineQuantizedTensor. - Add UIntxBitPackedTensor tensor subclass with from_hp(), dequantize(), and aten.linear/t/slice dispatch implementations - Add UIntxWeightOnlyConfig and handler in prototype quant_api - Export UIntxWeightOnlyConfig from public quantization API - Add comprehensive tests (4-bit, 8-bit, dynamic, slice, non-standard shapes) - Old GemliteUIntXWeightOnlyConfig remains as deprecated alias [ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/4079
Note: Links to docs will display an error until the docs builds have been completed. ⏳ 11 Pending, 1 Unrelated FailureAs of commit 4e57a4b with merge base ab4a336 ( FLAKY - The following job failed but was likely due to flakiness present on trunk:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
jerryzh168
added a commit
that referenced
this pull request
Mar 13, 2026
…fig) Replace the old AQT-based GemliteUIntXWeightOnlyConfig with a new v2 UIntxWeightOnlyConfig backed by UIntxBitPackedTensor(TorchAOBaseTensor), removing the dependency on AffineQuantizedTensor. - Add UIntxBitPackedTensor tensor subclass with from_hp(), dequantize(), and aten.linear/t/slice dispatch implementations - Add UIntxWeightOnlyConfig and handler in prototype quant_api - Export UIntxWeightOnlyConfig from public quantization API - Add comprehensive tests (4-bit, 8-bit, dynamic, slice, non-standard shapes) - Old GemliteUIntXWeightOnlyConfig remains as deprecated alias ghstack-source-id: cce3e06 Pull Request resolved: #4079
meta-cla
Bot
added
the
CLA Signed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Stack from ghstack (oldest at bottom):
Replace the old AQT-based GemliteUIntXWeightOnlyConfig with a new v2
UIntxWeightOnlyConfig backed by UIntxBitPackedTensor(TorchAOBaseTensor),
removing the dependency on AffineQuantizedTensor.
and aten.linear/t/slice dispatch implementations